Skip to content

accelerate-data/vibedata-official

Repository files navigation

Accelerate Data

The official skill library and plugin marketplace for VibeData — the agentic coordination layer for data platforms.

Organizational data engineering standards, encoded and enforced.

Why This Exists

Data teams lose weeks to coordination tax — intent lost between role switches, fix patterns never captured, institutional knowledge trapped in people's heads. Current tools address individual slices (ingestion, transformation, observability) without connecting them.

This repository is the distribution layer for the standards that customize VibeData data engineering agents. Skills encode organizational rules for ingestion, extraction, source-system handling, platform conventions, and business logic so agents build, deploy, and operate pipelines with your team's accumulated knowledge — not generic defaults.

The same bundles can also be installed directly in Claude Code as standalone skills for practitioners who want those standards available outside VibeData.

Skills vs Plugins

Skills Plugins
What Markdown knowledge packages that encode organizational standards, source customization, and business rules Installable Claude Code packages that bundle skills for VibeData agents or standalone Claude Code use
How they work Loaded into VibeData data engineering agents to align behavior with your team's standards Installed from the marketplace as reusable capabilities
Example dbt naming conventions for Fabric, SCD2 snapshot patterns Skill bundles like vibedata-dbt-skills, or workflow plugins like ad-migration

Skills are the behavior-shaping knowledge. Plugins are the distribution unit that makes those standards available to VibeData agents and, when useful, to Claude Code directly.

Quick Start

Add the marketplace:

/plugin marketplace add accelerate-data/vibedata-plugins-official

Install the flagship plugin:

/plugin install ad-migration@vibedata-plugins-official

Then explore the rest of the catalog from the same marketplace. Skill bundles are intended for VibeData agent customization first, and can also be installed directly in Claude Code.

Codex

Add the Codex marketplace:

codex plugin marketplace add accelerate-data/vibedata-plugins-official

The Codex marketplace is declared in .agents/plugins/marketplace.json, and each bundled plugin exposes a .codex-plugin/plugin.json manifest.

Featured Plugin: ad-migration

ad-migration helps teams modernize legacy data warehouses into governed Fabric Lakehouse platforms.

The result is a cleaner, agent-ready foundation for VibeData workflows and more predictable AI outcomes.

It is designed for teams that need to:

  • assess and scope legacy SQL Server warehouse assets
  • profile source structures and migration candidates
  • generate tests and dbt models as part of a repeatable workflow
  • move from one-off migration efforts to governed, reviewable delivery

Install:

/plugin install ad-migration@vibedata-plugins-official

Available Plugins

Plugin What it does Best for Install
ad-migration Governed migration workflows for stored procedures, warehouse logic, and dbt model generation Legacy warehouse to Fabric Lakehouse modernization /plugin install ad-migration@vibedata-plugins-official
fabric-cli Operational workflows for Fabric workspaces, notebooks, pipelines, and OneLake Fabric platform operations and environment control /plugin install fabric-cli@vibedata-plugins-official
fabric-semantic-model TMDL semantic model design, validation, and DAX measure authoring Fabric and Power BI semantic model delivery /plugin install fabric-semantic-model@vibedata-plugins-official
vibedata-dbt-skills Customizes dbt agents with Fabric modeling, snapshot, semantic layer, and Elementary quality standards dbt project delivery on Microsoft Fabric /plugin install vibedata-dbt-skills@vibedata-plugins-official
vibedata-dlt-skills Customizes ingestion agents with dlt REST API standards for ADLS Gen2, OneLake, and Fabric lakehouses API ingestion pipelines feeding dbt /plugin install vibedata-dlt-skills@vibedata-plugins-official
vibedata-domain-skills Customizes modeling agents with business rules, including revenue recognition Domain-aware marts and accounting logic /plugin install vibedata-domain-skills@vibedata-plugins-official
vibedata-ingestion-skills Customizes extraction agents with source-system standards, including Salesforce extraction CRM/source-system extraction patterns /plugin install vibedata-ingestion-skills@vibedata-plugins-official

Skill Bundle Contents

Skills are distributed through Claude plugins. Install the bundle that matches the VibeData agent behavior you want to customize, or install it directly in Claude Code for standalone use.

vibedata-dbt-skills

Skill What it encodes
dbt-fabric-patterns Practitioner-level dbt patterns for Microsoft Fabric
dbt-semantic-layer Semantic models and MetricFlow metrics in dbt on Microsoft Fabric
dbt-snapshot-scd2 SCD Type 2 snapshot implementation in dbt on Microsoft Fabric
elementary-data-quality Elementary anomaly detection configuration for dbt on Microsoft Fabric

vibedata-dlt-skills

Skill What it encodes
dlt-rest-api-connector dlt REST API pipelines to ADLS Gen2 and OneLake

vibedata-domain-skills

Skill What it encodes
modeling-revenue-recognition Revenue recognition mapped to dbt medallion architecture on Microsoft Fabric

vibedata-ingestion-skills

Skill What it encodes
salesforce-extraction Salesforce data extraction via dlt into dbt on Microsoft Fabric

How Skills Work in VibeData

Skills plug into three surfaces of the VibeData agentic workflow:

  • Build — When a practitioner describes business intent, agents draw on skills to validate domain fit and generate code that follows your team's conventions.
  • Deploy — When a PR is opened, deploy agents use skills to run context-informed quality gates — checking documentation, code quality, test coverage, and data quality.
  • Operate — When a production incident is resolved, fix patterns are captured as new skills and tests, preventing recurrence on future deployments.

Every pipeline built and every incident resolved makes the skill library more valuable. This is the improvement flywheel: accumulated domain expertise that compounds over time.

Coming Soon

Plugin Planned focus
fabric-ontology Business ontology and semantic layer design workflows for governed analytics platforms

Contributing

Contributions are welcome, but this repository is curated. If you want to contribute a skill or plugin, or improve an existing one:

  1. Open an issue or pull request with the proposed change.
  2. Keep the README and marketplace entry aligned with the actual purpose.
  3. Follow the repository conventions in CLAUDE.md.

License

Elastic License 2.0. See LICENSE.

About

No description, website, or topics provided.

Resources

License

Contributing

Security policy

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages